In the rapidly evolving world of Generative AI, three prominent Asian countries, China, Japan, and South Korea, continue to aggressively power their innovation engine. Each country presents a unique landscape with its own strengths, challenges, and aspirations, shaping the trajectory of Generative AI development in their respective regions. From China's robust investment and burgeoning startup ecosystem to Japan's efforts to catch up in large language model (LLM) development, and South Korea's determination to become a top AI powerhouse, these countries offer a fascinating glimpse into the potential of Generative AI.
Let us delve into these geographical markets, and uncover the distinct pathways these nations are forging in the exciting realm of Generative AI.
China
China is leading the global surge in investment in generative AI, with significant funding raised by Chinese startups in the first half of this year. According to Zhidongxi, out of 51 generative AI start-ups worldwide that raised about 100 billion yuan (US$13.8 billion) in funding in the first half of 2023, 22 startups were from China, the US had 21, and the UK had 4. Although China outperformed the US in terms of the number of startups getting funded, US-based startups got more funding than their Chinese counterparts. Some examples of Chinese companies funded in H1 2023 include Light Year, DeepLang AI, and MiniMax, which were all funded by Tencent. The biggest funding of 1.6 billion yuan was raised by Light Year, which later got acquired by Meituan.
Chinese organizations, including tech giants Alibaba and Tencent, have launched at least 79 large language models (LLMs) domestically over the past three years, contributing to the country's advancements in generative AI. This demonstrates the significant role played by both Chinese companies and international powerhouses like OpenAI, Microsoft, Google, and Meta in propelling LLM advancements in China.
China's private sector has made significant progress in generative AI. Baidu, known for its leading search engine, is developing the "Ernie Bot," an advanced chatbot capable of language understanding, generation, and text-to-image generation. Alibaba's DAMO Academy is internally testing its own ChatGPT-style chatbot, while JD Cloud, the cloud computing branch of JD.com, is launching ChatJD, an AI platform for various sectors, including retail and finance. Qihoo 360, an internet security company, plans to release a demo version of a generative AI tool, and Kuaishou, a short video-sharing social media company, is engaged in LLM research and its applications.
However, China's tech industry is grappling with a stricter regulatory environment surrounding AI. Censorship policies and the responsibility of platform providers for training algorithms and implementing censorship measures present challenges. Moreover, access to powerful chipsets required for training large AI models is restricted due to international trade limitations.
Despite these challenges, the Chinese Generative AI landscape presents significant growth opportunities. Chinese billionaires and tech industry leaders are investing heavily in AI technologies, aiming to compete globally and secure a share of the market. With an estimated global market opportunity of $800 billion over the next decade, the race for AI dominance in China is in its early stages, and stakeholders are determined to seize the potential for innovation and success.
Japan
Japan's Generative AI landscape is experiencing both opportunities and challenges. According to a survey conducted by Teikoku Databank, 61% of Japanese firms have a positive stance on utilizing generative AI in their operations. While only 9.1% of companies are currently using generative AI, 52% are considering its adoption. However, concerns about data breaches have led 23.3% of companies to be reluctant, which could significantly impact the uptake of generative AI in Japan.
When it comes to LLM development, Japan lags behind the US, China, and Europe in terms of scale and speed. This is attributed to the country's comparative shortcomings in deep learning and extensive software development. Japan is projected to face a deficit of 789,000 software engineers by 2030, according to the Ministry of Economy Trade and Industry. Hardware challenges also exist as LLMs require AI supercomputers for training, but no private company in Japan possesses a "world-class machine" with such capabilities.
To address these gaps, efforts are underway to leverage Japan's supercomputing capabilities. Tokyo Institute of Technology and Tohoku University are collaborating with Fujitsu and Riken, the developers of Fugaku supercomputer, to develop LLMs based primarily on Japanese data. The Japanese government is also investing 6.8 billion yen ($48.2 million) to build a new supercomputer in Hokkaido, which would specialize in LLM training. These initiatives aim to boost Japan's development of generative AI and promote research collaboration.
Japanese tech companies, including SoftBank and NTT, are also actively pursuing generative AI development to enhance Japan's standing in the field. SoftBank's mobile arm plans to take an offensive posture, and develop its own generative AI platform, while NTT aims to create a specialized LLM focused on the Japanese language and culture. These private sector efforts mark Japan's initial strides in the generative AI space, requiring substantial capital investment and a skilled workforce in natural language processing and high-performance computing.
Increased participation in generative AI development aligns with a positive attitude toward AI adoption in other sectors in Japan. Hitachi has established a generative AI center to promote safe and effective use of the technology. With the expertise of data scientists and AI researchers, the center will formulate guidelines to mitigate risks associated with generative AI. The Japanese government is also open to considering the adoption of AI technology like ChatGPT, provided cybersecurity and privacy concerns are addressed.
Hence, although Japan faces challenges in catching up with global leaders in generative AI, the country is making focused efforts to bridge the gap. By leveraging supercomputing capabilities, private sector initiatives, and a positive attitude toward AI adoption, Japan aims to strengthen its position in the generative AI landscape and promote safe and effective use of the technology.
South Korea
South Korea is making significant strides in the field of Generative AI and aims to become one of the world's top three AI powerhouses by 2027, trailing behind the U.S. and China, according to the Ministry of Science and ICT (MSIT), South Korea. The country recognizes the importance of memory semiconductor technology in supporting large language models like ChatGPT, which rely on high-performance memory chips to remember past conversations and user preferences.
To strengthen its capabilities in the memory semiconductor field, South Korea's leading dynamic random-access memory chipmakers, Samsung Electronics and SK Hynix, are actively investing in AI research and development. Samsung Electronics plans to invest a staggering 300 trillion Korean won ($228 billion) in a new semiconductor facility, while OpenAI's CEO, Sam Altman, has encouraged South Korea to take the lead in AI chip production.
The South Korean government is also heavily investing in AI. In 2022, the MSIT announced the deployment of 1.02 trillion won ($786 million) in funding for AI semiconductor research and development over the next five years. Additionally, substantial funds, amounting to 826.2 billion won, will be allocated through 2030 to build high-end chips and collaborate with startups through data centers. This strategic investment aims to enhance the economic and industrial value of AI semiconductors, leveraging South Korea's strengths in the memory chip and foundry sector.
South Korea is not only investing in established chipmakers but also fostering innovation through startups. One such example is Rebellions, a South Korean AI chip design startup that claims to have developed a chip that surpasses performance standards, and outperforms Nvidia's equivalent GPUs by over three times. As the government aims to boost local companies, Rebellions is actively vying for government contracts, competing with U.S. chip giants.
With its focus on AI semiconductor development, strategic investments, and support for startups, South Korea is positioning itself as a formidable player in the Generative AI landscape. By leveraging its memory chip expertise and fostering technological advancements, the country aims to establish itself as a leading force in Generative AI research and development.
Conclusion
The Generative AI landscapes in China, Japan, and South Korea demonstrate unique characteristics and varying levels of advancement in their respective countries. Each country presents some unique opportunities and challenges in the field of generative AI. While China has demonstrated significant startup growth, Japan and South Korea are leveraging their technological expertise and making strategic investments to bridge the gaps. As the global AI race continues, these countries' advancements in generative AI will shape the future of Gen AI innovation, impacting industries and society on a global scale.